Personalization Exercise Recommendation Based on Cognitive Diagnosis

Abstract

Online education has become an important part of education services. Faced with the vast online learning resources, how to provide students with personalized learning resources and achieve the goal of precise teaching is an important part of the intelligent teaching system. The premise of personalized recommendation is accurate personalized cognitive diagnosis and accurate analysis of each student's mastery of knowledge. Since cognitive diagnosis can diagnose and evaluate individual knowledge structure or cognitive process, combined with knowledge space theory, this paper designs a personalized exercise recommendation method based on cognitive diagnosis. The sequential dependency between course knowledge points has an impact on personalized exercise recommendation. Therefore, firstly, according to the knowledge space theory, the relationship graph between the knowledge concepts of the course is constructed, that is, the course knowledge topology map; Students conduct cognitive diagnosis and analyze the knowledge mastery status of each student; finally, the knowledge structure of each student is visualized on the knowledge map, and personalized test questions are recommended based on students' knowledge structure, question difficulty and course knowledge map. While improving the accuracy and reliability of the recommended test items, the interpretability of test item recommendation is guaranteed. The experimental comparison proves the accuracy and interpretability of the method in test item recommendation.